Reproduce “Entity Alignment For Knowledge Graphs: Progress, Challenges, and Empirical Studies” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Deepak Chaurasiya, Anil Surisetty, Nitish Kumar, Alok Singh, Vikrant Dey, Aakarsh Malhotra, Gaurav Dhama, Ankur Arora
The authors audit embedding-based entity alignment methods and introduce a low-name-bias test set. On this test set, Hits@1 falls from 77.93% to 61.10% for HGCN and from 75.17% to 61.10% for RDGCN; AttrE falls from 45.25% to 38.77%. Because sampling also changes dataset properties, these differences show sensitivity to the constructed test set, not that names alone caused them.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.